Increased frequency and mortality in persons with neurological disorders during COVID-19

Abstract

Determining the frequency and outcomes of neurological disorders associated with coronavirus disease 2019 (COVID-19) is imperative for understanding risks and for recognition of emerging neurological disorders. We investigated the susceptibility and impact of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection among persons with premorbid neurological disorders, in addition to the post-infection incidence of neurological sequelae, in a case–control population-based cohort.

Using health service data collected between 1 March 2020 and 30 June 2021, we constructed a cohort of SARS-CoV-2 RNA-positive (n = 177 892) and -negative (n = 177 800) adults who were age, sex and comorbidity matched and underwent RT-PCR testing at similar times. COVID-19-associated mortality rates were examined within the cohort. Neurological sequelae were analysed during the acute (<3 months) and the post-acute (3–9 months) phases post-infection.

The risk of death was significantly greater in the SARS-CoV-2 RNA-positive (2140 per 100 000 person years) compared with RNA-negative (922 per 100 000 person years) over a follow-up of 9 months, particularly amongst those with premorbid neurological disorders: adjusted odds ratios (95% confidence interval) in persons with a prior history of parkinsonism, 1.65 (1.15–2.37); dementia, 1.30 (1.11–1.52); seizures, 1.91 (1.26–2.87); encephalopathy, 1.82 (1.02–3.23); and stroke, 1.74 (1.05–2.86). There was also a significantly increased risk for diagnosis of new neurological sequelae during the acute time phase after COVID-19, including encephalopathy, 2.0 (1.10–3.64); dementia, 1.36 (1.07–1.73); seizure, 1.77 (1.22–2.56); and brain fog, 1.96 (1.20–3.20). These risks persisted into the post-acute phase after COVID-19, during which inflammatory myopathy (2.57, 1.07–6.15) and coma (1.87, 1.22–2.87) also became significantly increased.

Thus, persons with SARS-CoV-2 infection and premorbid neurological disorders are at greater risk of death, and SARS-CoV-2 infection was complicated by increased risk of new-onset neurological disorders in both the acute and post-acute phases of COVID-19.

Marsters et al. studied the frequency and outcomes of neurological disease associated with COVID-19 in over 355 000 people and determined that individuals with pre-existing neurological disorders were at greater risk of death. SARS-CoV-2 infection also increased the risk of new-onset neurological disorders post-infection that contribute to long COVID.

Introduction

Recent data indicate that severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) genome is present in the nervous system during the acute and post-acute phases after initial infection.¹ Moreover, COVID-19 also contributes to the development of new-onset nervous system disorders.² Several neurological disorders can confer greater susceptibility to the effects of COVID-19: worsened mortality rates are reported outcomes for persons with Parkinson’s disease,³ dementia^4-6 and epilepsy.⁷ The onset of new neurological disease presenting in association with SARS-CoV-2 infection in hospitalized persons is associated with an increased burden of mortality and morbidity, even after the acute stages of viral infection.^8-11 To date, studies reporting incident neurological sequelae after SARS-CoV-2 infection have focused solely on hospitalized patients or have lacked control groups or have reported neurological disorders at varying time points after SARS-CoV-2 diagnosis. Of these studies, many show a high incidence of headache and encephalopathy/cognitive dysfunction in both the acute and chronic time periods after SARS-CoV-2 infection, with reports to a lesser extent of paresthesia, movement disorders, cerebrovascular disease, myopathies, neuropathies, Guillain–Barré syndrome, encephalitis and seizures.^8,10,12-15 Large studies investigating the risk of COVID-19-associated neurological outcomes have used comparisons with other viral illnesses, historical cohorts or SARS-CoV-2 negative controls, albeit not time or geographically matched, which report increased risk of neurological sequelae spanning variable periods post-infection.^16-23 More recently, studies have focused on post-acute sequelae of COVID-19 (PASC) or ‘long COVID’ (defined as >3 months after the initial infection), which exhibits a range of symptoms that involve multiple organs and systems.²⁴ Neurological disorders constitute a substantial component of PASC symptomatology and can include neurocognitive deficits, mood disorders, pain, autonomic dysfunctions, dysgeusia and anosmia.²⁵ Two recent studies show a higher risk of developing encephalopathy, neurocognitive dysfunction, neuropathy, anxiety, fatigue, stroke, headache, seizure and movement disorders after the acute phase of SARS-CoV-2 infection when compared with SARS-CoV-2-negative or historical controls,^17,22 although other studies did not observe increased risks.^18,23

In the present case–control population-based study, we examined the effects of SARS-CoV-2 infections on neurological disorders by assessing the impact of infection on ensuing mortality in persons with premorbid neurological disorders and on the incidence of onset neurological disorders diagnosed after PCR confirmation of SARS-CoV-2 infection. Defining the incidence and outcomes of neurological disorders in persons with COVID-19 is imperative for understanding risk status and for recognition of emerging post-acute sequelae of COVID-19-associated neurological syndromes.

Materials and methods

Population

This population-based retrospective cohort study was conducted on all 3.2 million adults residing in the province of Alberta in Canada, and our cases were those with SARS-CoV-2 RNA detection based on RT-PCR test results between 1 March 2020 and 30 June 2021. RT-PCR tests were collected from all settings (outpatient clinics, inpatient settings and community testing centres). For persons with multiple positive SARS-CoV-2 PCR tests, only the first positive test result was considered. Controls were persons within Alberta and a first negative SARS-CoV-2 PCR test matched by test date within a month during the same month that were matched by age, sex and Charlson score, and by construction, given that we started with the PCR-positive cases, they did not have a subsequent positive test during the study period (Fig. 1). The study was approved by the Institutional Ethics Review board at the University of Alberta (Study ID: Pro00116826).

Figure 1.

Flow chart of cohort construction based on all SARS-CoV-2 RT-PCR tests done between 13 March 2020 and 30 June 2021. Patients aged <18 years old were excluded. Patient ascertainment was based on Alberta residency with a valid Alberta Personal Healthcare Number (PHN). Cases consisted of all those within the timespan who were SARS-CoV-2 RNA-positive ([+]). Controls were matched for age, sex and Charlson score and were found to be SARS-CoV-2 RNA-negative ([−]) throughout the study period. A nested subgroup analysis was performed from within this cohort of SARS-CoV-2 RNA[+] persons with premorbid neurological disease diagnosed 2 years prior to PCR testing along with a matched group without a premorbid neurological disease and were age, sex and Charlson score matched.

Setting

Alberta has a publicly funded, universal-access healthcare system, free of charge at the point of care, with capture of all interactions with the healthcare system. Six provincial healthcare service databases were used for this study, and patients were linked across databases via their encrypted unique health identifier number, including: (i) the Discharge Abstract Database (DAD) records all acute care hospitalizations, with ≤25 diagnoses and admission/discharge dates; (ii) the Ambulatory Care Database captures all emergency department visits and hospital-based physician office visits with ≤10 diagnostic codes; (iii) the Healthcare Provider Claims Database captures all physician visits (including those shadow-billed by salaried physicians) with up to three diagnoses; (iv) Alberta Precision Laboratory data stores the results of SARS-CoV-2 tests; (v) the Pharmaceutical Information Network (PIN) database records all medication dispensations at community-based pharmacies in Alberta; and (vi) the Alberta Health Care Insurance Program Central Stakeholder Registry holds demographic data, including home address (used for defining socioeconomic status) and date of emigration or death. The Data and Research Services Platform of the Alberta Strategy for Patient Oriented Research Support Unit (AbSPORU) conducted data extraction and statistical analyses.

Comorbidities

Previously validated International Classification of Diseases (ICD)-9- and ICD-10-based case definitions were used from the hospital admission, emergency department assessments and provider claims datasets²⁶ in the 2 years prior to (and including) the index date. Premorbid neurological disease diagnoses were identified from our database using the ICD-10 codes (Supplementary Table 1) based on an established diagnosis within the 2 years prior to the SARS-CoV-2 RT-PCR testing.

Outcomes

The acute phase was defined as between 7 days and 3 months after the SARS-CoV-2 PCR test, and the post-acute phase (overlapping with ‘long COVID’, also termed PASC) was defined as between 3 and 9 months after the first SARS-CoV-2 PCR test. The cohort was analysed by emergency department assessments or hospital admissions during the first month of index date, or died during follow-up, whereas persons not assessed in the emergency department or admitted to hospital were deemed as home assessments. Deaths for mortality rate calculation were assessed for ≤9 months after the index PCR test date. Incident neurological disorders post-PCR test were identified from our database using the ICD-10 codes concentrating on neurological disorders (Supplementary Table 1). The date of first emergency department visit or hospital admission for the neurological disorder, as determined from recorded listed diagnoses, was based on chart review at discharge and used to determine incidence.

Statistical analysis

Demographic and clinical data and comorbidities, including premorbid neurological conditions, were analysed by comparing persons who tested PCR-positive or -negative for SARS-CoV-2. To compare the risks with premorbid neurological disorders, the odds ratios (ORs) for death were calculated for persons with a SARS-CoV-2 (RNA) PCR test and a pre-existing neurological disorder. A subgroup analysis was performed within the SARS-CoV-2 RNA-positive group, for which age-, sex- and Charlson score-matched groups with and without premorbid neurological disorders were matched to determine the adjusted OR (aOR) of death for individual neurological disorders within the SARS-CoV-2 PCR-positive group. Kaplan–Meier curves were generated to examine the incidence and timing of new neurological disorders from 7 days to 9 months after the index PCR test. A separate analysis was performed to investigate ORs for incident neurological disorders in both the subacute phase (7 days to 3 months) and the post-acute phase (3–­9 months). All statistical analyses were done using SAS v.9.4 (Cary, NC, USA), and figures were generated using R v.4.1.2 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Differential survival within the case–control cohort depending on infection status

This study included adult persons who were PCR-positive for SARS-CoV-2 RNA in any setting (n = 177 892), together with a control group who were PCR-negative for SARS-CoV-2 RNA on tests done at the same time and were age, sex and comorbidity matched (n = 177 800) (Table 1). Material and social deprivation scores and Charlson Comorbidity Index scores were significantly greater in the SARS-CoV-2 PCR-positive group (both P < 0.0001). SARS-CoV-2 RNA-negative (control) subjects were more likely to have been PCR tested and diagnosed with COVID-19 during the study period as a home assessment, rather than diagnosed upon visit to the emergency department or hospital or death (176 594 at home versus 1251 emergency department/hospital) when compared with the SARS-CoV-2 RNA-positive group (159 113 at home versus 18 779 emergency department/hospital). There was a lower mortality rate among the SARS-CoV-2 RNA-negative controls compared with the SARS-CoV-2 RNA-positive cases [1229 (2140 per 100 000 person years) versus 2856 (922 per 100 000 person years) deaths, respectively, over a median follow-up of 9 months]. SARS-CoV-2 infection and the associated risk of mortality was assessed among persons with premorbid neurological disorders diagnosed ≥2 years prior to the initial PCR test for SARS-CoV-2 (Supplementary Table 2).

Table 1.

Demographic features of the SARS-CoV-2 RNA-positive and matched SARS-CoV-2 RNA-negative groups as a case–control cohort

	Controls, SARS-CoV-2 RNA-negative	Cases, SARS-CoV-2 RNA-positive
Subjects, n	177 800	177 892
Sex
Female	88 369 (49.7%)	88 484 (49.7%)
Male	89 414 (50.3%)	89 408 (50.3%)
Age, years
Mean (SD)	42.19 (16.84)	42.69 (16.86)
Median (interquartile range)	39 (29, 53)	40 (29, 53)
Age group, n (%)
18–29 years	47 023 (26.4%)	57 927 (32.6%)
30–39 years	42 073 (23.7%)	37 929 (21.3%)
40–49 years	34 980 (19.7%)	31 841 (17.9%)
50–59 years	26 012 (14.6%)	23 838 (13.4%)
60–69 years	15 048 (8.5%)	14 106 (7.9%)
70–79 years	6503 (3.7%)	6057 (3.4%)
80+ years	6161 (3.5%)	6194 (3.5%)
Region, n (%)
Metropolitan	26 089 (14.7%)	107 793 (60.6%)
Regional	132 531 (74.5%)	49 255 (27.7%)
Rural	19 180 (10.8%)	20 844 (11.7%)
Combined material/social deprivation, n (%)
1 (least deprived)	29 727 (16.7%)	24 846 (14.0%)
2	36 296 (20.4%)	32 893 (18.5%)
3	34 772 (19.6%)	38 650 (21.7%)
4	28 625 (16.1%)	30 233 (17.0%)
5 (most deprived)	34 341 (19.3%)	40 558 (22.8%)
Charlson score
Mean (SD)	0.10 (0.59)	0.18 (0.75)
Deaths (median 9.0 months of follow-up)	1229	2856

Groups are matched by sex, age, and Charlson score. Numerical values are given as number (n) and percentage (%) unless otherwise specified.

Impact of premorbid neurological disorder on SARS-CoV-2 infection

When compared with the SARS-CoV-2 RNA-negative control group, persons with SARS-CoV-2 infection showed higher odds of having premorbid parkinsonism [OR 1.53, 95% confidence interval (CI) 1.21–1.94], inflammatory myopathy (OR 1.40, 95% CI 1.00–1.96), encephalopathy (OR 1.89, 95% CI 1.47–2.43), dementia (OR 2.38, 95% CI 2.16–2.61), seizure/epilepsy (OR 1.35, 95% CI 1.18–1.55), headache disorder (OR 1.12, 95% CI 1.04–1.21), ischaemic stroke (OR 1.59, 95% CI 1.36–1.86), haemorrhagic stroke (OR 1.53, 95% CI 1.18–2.00), cerebrovascular disease (OR 1.57, 95% CI 1.40–1.77) and myelitis (OR 2.08, 95% CI 1.05–4.14) (Supplementary Table 2). An increased neurological disorder-associated prevalence in the SARS-CoV-2 RNA-positive group was not observed for persons with demyelinating disease, myasthenia gravis, congenital or hereditary myopathies, inflammatory polyneuropathies or hereditary neuropathies (Supplementary Table 2). The risk of mortality among the SARS-CoV-2 RNA-positive group showed a higher unadjusted odds of death if they had a premorbid neurological diagnosis of parkinsonism (OR 1.88, 95% CI 1.05–3.38), encephalopathy (OR 2.75, 95% CI 1.02–7.38), dementia (OR 1.47, 95% CI 1.19–1.83), headache (OR 4.16, 95% CI 1.19–14.5), ischaemic stroke (OR 1.67, 95% CI 1.08–2.58) and cerebrovascular disease (OR 1.84, 95% CI 1.29–2.62) compared with SARS-CoV-2 RNA-negative persons without these premorbid conditions (Fig. 2 and Supplementary Table 3). This finding was not evident among persons with a premorbid demyelinating disease, myasthenia gravis, inflammatory myopathies, seizure/epilepsy or haemorrhagic stroke. An OR could not be calculated for Huntington’s disease, congenital or hereditary myopathies, inflammatory polyneuropathies, motor neuron disease and myelitis because of their low prevalence in our cohort.

Figure 2.

Unadjusted odds ratios of death among SARS-CoV-2 RNA-positive persons compared with RNA-negative persons. Based on premorbid neurological disorders displayed as unadjusted odds ratios and 95% confidence intervals.

Effect of premorbid neurological disorder on survival after COVID-19

Given the range of comorbidities that accompanied these neurological disorders, a nested cohort of SARS-CoV-2 PCR-positive persons, adjusting for age, sex and Charlson score, was analysed to investigate whether a premorbid neurological disorder was associated with higher odds of death in persons with SARS-CoV-2 infection. This adjusted subgroup analysis was performed on persons who were SARS-CoV-2 RNA-positive and diagnosed with a premorbid neurological disorder (n = 4667) together with a group of matched controls who were SARS-CoV-2 RNA-positive without a premorbid neurological disorder (n = 4667) (Table 2). A logistic regression analysis revealed greater mortality risk in persons with SARS-CoV-2 infection (Fig. 3) and premorbid diagnoses of parkinsonism (aOR 1.65, 95% CI 1.15–2.37), encephalopathy (aOR 1.82, 95% CI 1.02–3.23), dementia (aOR 1.30, 95% CI 1.11–1.52), seizure/epilepsy (aOR 1.91, 95% CI 1.26–1.52) and haemorrhagic stroke (aOR 1.74, 95% CI 1.05–2.86) when compared with RNA-positive individuals without premorbid neurological diagnoses (Table 2). In this nested analysis, we did not detect differences in mortality risk for persons with SARS-CoV-2 infection and premorbid diagnoses of Huntington’s disease, demyelinating diseases, myasthenia gravis, congenital or hereditary myopathies, inflammatory myopathies, headache, ischaemic stroke, cerebrovascular disease, motor neuron disease or myelitis in comparison to persons with SARS-CoV-2 infection but without these neurological disorders (Supplementary Table 4).

Table 2.

Mortality and demographic features for SARS-CoV-2 RNA-positive persons with or without premorbid neurological disorder

	SARS-CoV-2 RNA-positive without neurological disorders		SARS-CoV-2 RNA-positive with neurological disorders
	Deaths	Total	Deaths	Total
Patients, n (%)	495	4667	606	4667
Sex, n (%)
Female	235 (47.5%)	2829 (60.6%)	270 (44.6%)	2829 (60.6%)
Male	260 (52.5%)	1838 (39.4%)	336 (55.4%)	1838 (39.4%)
Age, years
Mean (SD)	83.93 (10.86)	58.80 (23.45)	82.28 (12.14)	58.92 (23.60)
Median (interquartile range)	86 (80, 91)	58 (38, 81)	85 (77, 90)	58 (38, 81)
Age group, n (%)
18–29 years	23 (4.6%)	969 (20.8%)	28 (4.6%)	978 (21.0%)
30–39 years	3 (0.6%)	575 (12.3%)	3 (0.5%)	575 (12.3%)
40–49 years	3 (0.6%)	541 (11.6%)	5 (0.8%)	536 (11.5%)
50–59 years	8 (1.6%)	446 (9.6%)	19 (3.1%)	441 (9.4%)
60–69 years	24 (4.8%)	396 (8.5%)	41 (6.8%)	381 (8.2%)
70–79 years	67 (13.5%)	505 (10.8%)	106 (17.5%)	498 (10.7%)
80+ years	367 (74.1%)	1235 (26.5%)	404 (66.7%)	1258 (27.0%)
Region, n (%)
Metropolitan	304 (61.4%)	2769 (59.3%)	394 (65.0%)	2553 (54.7%)
Regional	109 (22.0%)	1198 (25.7%)	145 (23.9%)	1287 (27.6%)
Rural	82 (16.6%)	700 (15.0%)	67 (11.1%)	827 (17.7%)
Combined material/social deprivation, n (%)
1	53 (10.7%)	554 (11.9%)	85 (14.0%)	538 (11.5%)
2	61 (12.3%)	750 (16.1%)	95 (15.7%)	661 (14.2%)
3	97 (19.6%)	951 (20.4%)	96 (15.8%)	920 (19.7%)
4	98 (19.8%)	798 (17.1%)	89 (14.7%)	809 (17.3%)
5	120 (24.2%)	1156 (24.8%)	145 (23.9%)	1279 (27.4%)
Charlson score
Mean (SD)	2.73 (2.00)	1.28 (1.72)	2.47 (2.03)	1.28 (1.72)
Median (interquartile range)	2 (1, 4)	1 (0, 2)	2 (1, 3)	1 (0, 2)
Neurological disorder, n (%)
Parkinsonism	–	–	52 (8.6%)	172 (3.7%)
Huntington’s disease	–	–	1 (0.2%)	7 (0.1%)
Demyelinating disease	–	–	5 (0.8%)	104 (2.2%)
Myasthenia gravis	–	–	2 (0.3%)	15 (0.3%)
Congenital or hereditary myopathy	–	–	1 (0.2%)	10 (0.2%)
Inflammatory myopathy	–	–	6 (1.0%)	78 (1.7%)
Encephalopathy	–	–	21 (3.5%)	154 (3.3%)
Dementia	–	–	422 (69.6%)	1413 (30.3%)
Inflammatory polyneuropathy	–	–	0 (0.0%)	16 (0.3%)
Seizure/epilepsy	–	–	35 (5.8%)	466 (10.0%)
Ischaemic stroke	–	–	78 (12.9%)	395 (8.5%)
Haemorrhagic stroke	–	–	25 (4.1%)	134 (2.9%)
Cerebrovascular disease	–	–	123 (20.3%)	711 (15.2%)
Motor neuron disease	–	–	2 (0.3%)	6 (0.1%)
Hereditary neuropathy	–	–	0 (0.0%)	1 (0.0%)
Myelitis	–	–	3 (0.5%)	25 (0.5%)

Matched for sex, age and Charlson score. Numerical values are given as number (n) and percentage (%).

Figure 3.

Adjusted odds ratios of death among SARS-CoV-2 RNA-positive persons based on the nested analysis including the presence or absence of specified premorbid neurological disorders. Adjusted odds ratios using logistic regression and 95% confidence intervals are shown (adjusted for age, sex and Charlson score).

COVID-19-associated neurological disorder sequelae

To define the type and risks of new neurological disorders after SARS-CoV-2 infection, two post-infection phases were analysed. These periods were defined as an acute phase (from 7 days to 3 months after the index test date) (Fig. 4A) and a post-acute phase (3–9 months after the index test date) (Fig. 4B) after initial SARS-CoV-2 PCR test (Supplementary Table5). The acute phase showed higher OR of encephalopathy (OR 2.00, 95% CI 1.10–3.64), dementias (OR 1.36, 95% CI 1.07–1.73), seizure/epilepsy (OR 1.77, 95% CI 1.1.22–2.56) and brain fog (OR 1.96, 95% CI 1.20–3.20) in the SARS-CoV-2 RNA-positive group when compared with the SARS-CoV-2 RNA-negative group. Myelitis showed a higher frequency, although an OR could not be calculated because there were only five persons in SARS-CoV-2 RNA-positive group and no cases in the SARS-CoV-2 RNA-negative group. There were no differences in frequency between those with positive versus negative RNA tests for parkinsonism, demyelinating disease, myasthenia gravis, inflammatory myopathy/polyneuropathy, headache, ischaemic/haemorrhagic stroke, combined cerebrovascular disease, coma or ataxia in the acute phase (Supplementary Table 5). In the post-acute phase, corresponding to long COVID or PASC, the SARS-CoV-2 RNA-positive group showed a higher frequency of inflammatory myopathy (OR 2.57, 95% CI 1.07–6.15), encephalopathy (OR 2.06, 95% CI 1.15–3.68), dementia (OR 1.38, 95% CI 1.12–1.70), seizure/epilepsy (OR 1.37, 95% CI 1.05–1.79), brain fog (OR 1.64, 95% CI 1.12–2.41) and coma (OR 1.87, 95% CI 1.22–2.87) when compared with the SARS-CoV-2 RNA-negative group. No differences in frequency of new-onset neurological disorder were observed for parkinsonism, demyelinating disease, headache, ischaemic stroke, haemorrhagic stroke, combined cerebrovascular disease, abnormal involuntary movements or ataxia (Supplementary Table 5). ORs were not calculated for myasthenia gravis, inflammatory polyneuropathy and myelitis because of too few cases. To determine the frequency of new-onset neurological disorders over the combined acute and post-acute phases, Kaplan–Meier curves were generated to enable analyses of incidence over time from 7 days to 9 months after initial SARS-CoV-2 PCR testing. Kaplan–Meier curves showed that dementia, encephalopathy, seizure/epilepsy, brain fog and myelitis were significantly more likely to arise as neurological sequelae after a positive PCR test than a negative test (Fig. 5). In contrast, cerebrovascular disease (ischaemic and haemorrhagic stroke; Fig. 5), demyelinating disease, parkinsonism, myopathies, movement disorders and neuropathy showed similar incidence rates among subjects who were SARS-CoV-2 PCR-positive or -negative.

Figure 4.

Odds ratios of incident neurological disorders among SARS-CoV-2 RNA-positive relative to RNA-negative persons post-PCR testing. During (A) the acute phase (from 7 days to 3 months) and (B) the post-acute phase (3–9 months). Data are shown as unadjusted odds ratios with 95% confidence intervals.

Figure 5.

Kaplan–Meier curves depicting risk of developing a neurological disorder. Time line from 7 days to 9 months after SARS-CoV-2 PCR testing [RNA-negative in green (Controls) versus RNA-positive in magenta (Cases)]. Shaded areas represent the 95% confidence interval.

Discussion

To the best of our knowledge, the present study represents the largest case–control population-based study to have investigated survival outcomes after SARS-CoV-2 infection among persons with a prior neurological disorder and the frequency of new neurological sequelae in patients diagnosed in the community and not hospitalized during the acute and post-acute phases after COVID-19. These data highlight specific neurological disorders that develop as post-acute sequelae of COVID-19 (PASC phase) and underscore the adverse impacts of COVID-19 on the emergence of neurological diseases. The present findings also emphasize the effects of a systemic infectious disease on the incidence and mortality among persons affected by a range of neurological disorders.

In the present study, persons who were SARS-CoV-2 PCR-positive had significantly higher odds of death with a premorbid diagnosis of parkinsonism, encephalopathy, dementia, headache disorder and stroke (Fig. 2), which persisted after adjusting for age, sex and Charlson score (Fig. 3). Overall, these data emphasize an association between specific premorbid neurological disorders and dying from COVID-19. Although multiple factors contribute to this increased risk, it highlights a subpopulation of persons who were previously diagnosed with specific neurological disorders that might benefit further from risk-modifying strategies, such as vaccines or antiviral medications, or might need to be monitored more carefully after contracting COVID-19 to optimize clinical outcomes. Our findings are aligned with other studies that show an association of increased mortality rate during the COVID-19 pandemic in those with premorbid Parkinson’s disease,³ dementia,^4-6 epilepsy^5,7 and stroke⁵ when compared with pre-pandemic rates or using within-cohort comparisons.

We found a higher incidence of encephalopathy, dementia, seizure/epilepsy, brain fog and myelitis within the acute phase from 7 days to 3 months after a positive SARS-CoV-2 PCR test. In the post-acute phase (3–9 months), we found a higher incidence of inflammatory myopathy and coma, in addition to a continued increase in the incidence of dementia, seizure/epilepsy and brain fog. The broader implications of these findings are that the risks of being diagnosed with a neurological disease after SARS-CoV-2 infection continue for an extended period of time after infection and warrant heightened awareness of these disorders during medical follow-up after COVID-19. Other large cohort studies have also reported a post-COVID-19 increase in the incidence of encephalopathy,^{16,17,19,22,27} cognitive deficits or dementia,^19,22,23,27 and myopathy or neuromuscular junction disease,^19,22 while others have shown variable outcomes for increased incidence of stroke^{17-20,23,28-32} and epilepsy/seizures,^18,20,22,23 and no association with parkinsonism.^19,20,23 Nevertheless, these latter studies were heterogeneous in their methods, such as the type of neurological sequelae assessed, choice of observed time lines, selection of control groups, whether there was adjustment of variables, and resulting outcome measurements, making it difficult to compare results across studies.

Two large studies that examined the risk of stroke from 1 month to 1 year after index infection found conflicting results when comparing with COVID-19-negative contemporary controls.^22,23 Other studies using similar controls but different study periods show increased risk when the outcome was observed up to 30 days,³⁰ 1 year,²¹ within 3 weeks to 4 months,¹⁷ but not within 2 weeks to 6 months.¹⁸ Studies using other respiratory illnesses as a comparator also show variable results: one cohort study showed significant increased risk of any stroke up to 6 months that resolved by 2 years,^19,20 but another study with 90 days of observation did not report similar findings.³¹ Differences in risk across studies might be explained by differences in cohort variables such as age, selection criteria for cases and controls, or comorbidities within the included populations; variable span of time observed for outcomes; and whether hospitalized individuals or milder infections were included. Thus, drawing a definitive conclusion regarding the risk of incident stroke associated with COVID-19 remains a challenge. Our study included a large population-based cohort spanning an entire province supported by public healthcare, which enables a centralized healthcare data-acquisition source for clinical and laboratory variables and included all viral RNA-positive patients and not only those presenting to emergency departments or hospitals, minimizing selection bias.

The present study is also the first to investigate the impact of COVID-19 on a wide range of neurological disorders in both the premorbid and post-infection stages. Our study was predicated on PCR confirmation of SARS-COV-2 infection, with centralized laboratory results ensuring that these data were dependable. Lastly, the cohort covers the first wave of the COVID-19 pandemic, before the emergence of variants of concern, before shortages of test kit supplies restricted the availability of testing, before the discovery of various COVID-19 therapies that mitigated disease severity (and might impact post-acute sequelae) and before broad vaccine availability.

The present study faced challenges that include limited clinical information available through health service data, although we did use previously validated ICD-based case definitions, albeit with limited clinical guidelines for diagnoses such as brain fog. We were also unable to take into consideration multiple neurological diagnoses in the same person. Furthermore, as with all COVID-19 studies, our study was subject to data identifying only those with COVID-19 who were tested initially, although there might have been persons with mild symptoms who did not undergo PCR testing or might have undergone PCR testing at later dates. Of note, the diagnosis of neurological sequelae was dependent on a visit to the emergency department or admission to hospital, hence those with milder symptoms might have been missed. Our study did not distinguish those subjects who required intensive care unit admission, which might herald a greater risk of neurological sequelae.³³ Given that the majority of persons who presented to hospital were tested for SARS-CoV-2 RNA in our study period, whether or not they had COVID-19 symptoms, this biased the data on incident disease sequelae within a few days of a SARS-CoV-2 PCR test to be increased in the SARS-CoV-2 RNA-negative (control) group. Another limitation was the possibility of misclassification bias (false-negative or false-positive RNA-PCR tests) or unknown confounders that made the cases both more likely to test positive for SARS-CoV-2 and increased their susceptibility to post-COVID-19 sequelae. Finally, our study can demonstrate associations of SARS-CoV-2 infection with neurological disease sequelae and mortality but cannot establish causation.

Conclusion

In summary, the present data indicated that persons with a premorbid diagnosis of parkinsonism, encephalopathy, dementia, seizure/epilepsy, stroke and myelitis were at higher risk of death after contracting SARS-CoV-2 infection. This information might aid in convincing higher-risk individuals to accept vaccination or seek medical attention (to obtain antiviral therapy) soon after infection. Furthermore, persons becoming infected with SARS-CoV-2 appear to be at greater risk of developing subsequent neurological sequelae, including encephalopathy, dementia, seizure/epilepsy, abnormal involuntary movements, inflammatory myopathy, brain fog, coma and myelitis, with risks extending well beyond the infectious and post-infectious period. Thus, COVID-19 continues to burden healthcare systems, and these findings highlight the impact of COVID-19 on specific neurological disorders.

Data availability

Anonymized data not published within this article will be made available by request from any qualified investigator.

Funding

The Alberta Strategy for Patient Oriented Research Unit is funded by the Canadian Institutes for Health Research (CIHR), Alberta Innovates, University Hospital Foundation and in-kind support from the Universities of Alberta and Calgary, Alberta Health Services and the Woman and Children’s Health Research Institute. J.A.B., G.Y.L. and C..P are funded by CIHR.

Competing interests

The authors report no competing interests.

Supplementary material

Supplementary material is available at Brain online.